CPU monitoring for F5

[Livius]

Hi all,

Our Viprions are experiencing CPU spikes of > 80% on a hourly basis.

We already have the # tmsh modify /sys db cpu.logthreshold value 80 set, but that only triggers an alert if we have >5min threshold crossed.

Is there any way to get more granular, to monitor and alert when for 5s intervals?

Thanks in advance!

[Gajji]

Yes I guess, it is possible to monitor and alert on CPU spikes for smaller intervals using the BIG-IP system's event system and scripting capabilities.

[aaperson]

Are you looking for diagnostics or monitoring?  5s is a very short interval for monitoring.   What needs to happen is shorter polling interval, but that 5m interval may be a limitation of the service.  It may take a 3rd party monitoring tool to give you greater frequency.